Openembedded Core Discussions
 help / color / mirror / Atom feed
* [PATCH] native.bbclass: Ensure native recipes have a deterministic baselib value
@ 2011-10-06 14:20 Richard Purdie
  2011-10-06 14:47 ` Mark Hatle
  2011-10-06 21:01 ` McClintock Matthew-B29882
  0 siblings, 2 replies; 4+ messages in thread
From: Richard Purdie @ 2011-10-06 14:20 UTC (permalink / raw)
  To: openembedded-core

Changes to baselib by specific machine configuration were resulting
in sstate cache invalidation, particularly in multilib configurations.

This patch ensures this doesn't happen and native sstate cache files
are reusable.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
---
diff --git a/meta/classes/native.bbclass b/meta/classes/native.bbclass
index 5e45aed..ba8b0bf 100644
--- a/meta/classes/native.bbclass
+++ b/meta/classes/native.bbclass
@@ -69,6 +69,8 @@ exec_prefix = "${STAGING_DIR_NATIVE}${prefix_native}"
 
 libdir = "${STAGING_DIR_NATIVE}${libdir_native}"
 
+baselib = "lib"
+
 # Libtool's default paths are correct for the native machine
 lt_cv_sys_lib_dlsearch_path_spec[unexport] = "1"
 





^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] native.bbclass: Ensure native recipes have a deterministic baselib value
  2011-10-06 14:20 [PATCH] native.bbclass: Ensure native recipes have a deterministic baselib value Richard Purdie
@ 2011-10-06 14:47 ` Mark Hatle
  2011-10-06 21:47   ` Richard Purdie
  2011-10-06 21:01 ` McClintock Matthew-B29882
  1 sibling, 1 reply; 4+ messages in thread
From: Mark Hatle @ 2011-10-06 14:47 UTC (permalink / raw)
  To: openembedded-core

On 10/6/11 9:20 AM, Richard Purdie wrote:
> Changes to baselib by specific machine configuration were resulting
> in sstate cache invalidation, particularly in multilib configurations.
> 
> This patch ensures this doesn't happen and native sstate cache files
> are reusable.

Likely throwing in a can of worms here, but for our existing (non-OE based)
work, we generally change the baselib to "lib64" on 64-bit machines.  Without
patching native.bbclass, is it possible to do that?

We do this because we ship both 32-bit and 64-bit host tooling for a variety of
host distributions, I expect we'll have to continue doing that as we transition
to an OE-based system.

> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
> ---
> diff --git a/meta/classes/native.bbclass b/meta/classes/native.bbclass
> index 5e45aed..ba8b0bf 100644
> --- a/meta/classes/native.bbclass
> +++ b/meta/classes/native.bbclass
> @@ -69,6 +69,8 @@ exec_prefix = "${STAGING_DIR_NATIVE}${prefix_native}"
>  
>  libdir = "${STAGING_DIR_NATIVE}${libdir_native}"
>  
> +baselib = "lib"
> +
>  # Libtool's default paths are correct for the native machine
>  lt_cv_sys_lib_dlsearch_path_spec[unexport] = "1"
>  
> 
> 
> 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core




^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] native.bbclass: Ensure native recipes have a deterministic baselib value
  2011-10-06 14:20 [PATCH] native.bbclass: Ensure native recipes have a deterministic baselib value Richard Purdie
  2011-10-06 14:47 ` Mark Hatle
@ 2011-10-06 21:01 ` McClintock Matthew-B29882
  1 sibling, 0 replies; 4+ messages in thread
From: McClintock Matthew-B29882 @ 2011-10-06 21:01 UTC (permalink / raw)
  To: Patches and discussions about the oe-core layer

On Thu, Oct 6, 2011 at 9:20 AM, Richard Purdie
<richard.purdie@linuxfoundation.org> wrote:
> Changes to baselib by specific machine configuration were resulting
> in sstate cache invalidation, particularly in multilib configurations.
>
> This patch ensures this doesn't happen and native sstate cache files
> are reusable.
>
> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

This fixes native sstate-cache (re)usage when enabling multilib.

-M



^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] native.bbclass: Ensure native recipes have a deterministic baselib value
  2011-10-06 14:47 ` Mark Hatle
@ 2011-10-06 21:47   ` Richard Purdie
  0 siblings, 0 replies; 4+ messages in thread
From: Richard Purdie @ 2011-10-06 21:47 UTC (permalink / raw)
  To: Patches and discussions about the oe-core layer

On Thu, 2011-10-06 at 09:47 -0500, Mark Hatle wrote:
> On 10/6/11 9:20 AM, Richard Purdie wrote:
> > Changes to baselib by specific machine configuration were resulting
> > in sstate cache invalidation, particularly in multilib configurations.
> > 
> > This patch ensures this doesn't happen and native sstate cache files
> > are reusable.
> 
> Likely throwing in a can of worms here, but for our existing (non-OE based)
> work, we generally change the baselib to "lib64" on 64-bit machines.  Without
> patching native.bbclass, is it possible to do that?

Yes, you could set this based on the build system architecture no
problem at all.

The reason for this change is to have a single value for native packages
which doesn't change based on MACHINE. That doesn't stop you changing it
based on something specific the native binaries themselves.

The native sysroots are different depending on the build system
architecture anyway so this actually might be a non-issue for you
anyway.

Cheers,

Richard




^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2011-10-06 21:53 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-10-06 14:20 [PATCH] native.bbclass: Ensure native recipes have a deterministic baselib value Richard Purdie
2011-10-06 14:47 ` Mark Hatle
2011-10-06 21:47   ` Richard Purdie
2011-10-06 21:01 ` McClintock Matthew-B29882

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox